What Is Acoustic Monitoring and How Is It Used?

Acoustic monitoring is the practice of using microphones, hydrophones, or vibration sensors to capture sound and then analyzing those recordings to learn something about the source. It spans an enormous range of applications: ecologists leave small recorders in forests to track bird populations, engineers attach sensors to wind-turbine gearboxes to catch early signs of bearing damage, and physicians are developing wearable stethoscopes that continuously listen to a patient’s lungs. The unifying idea is that sound carries information, and with the right hardware and analysis, you can extract that information remotely, continuously, and often at a fraction of the cost of sending a human to look.

Passive Versus Active Listening

In ecology, the term most people encounter is passive acoustic monitoring, or PAM. “Passive” means the recorder simply listens to whatever sounds the environment produces. It does not emit a signal of its own. A PAM device left in a wetland captures frog choruses, insect stridulations, wind, rain, and the occasional passing vehicle without disturbing any of them. By contrast, active acoustic methods do exist: sonar pings used to map the seafloor or ultrasonic pulses bounced off a material to check for internal cracks. But the passive approach has become the dominant paradigm for biodiversity research because it is cheap, non-invasive, and can run unattended for weeks or months.

A study comparing passive recorders with traditional human surveys in tropical savanna wetlands illustrates why. The passive devices detected all frog species found by human surveyors and consistently found more: roughly 14 species per site versus about 8 with active human searches. The advantage grew as the recorders covered a wider slice of the day and season, picking up species that call only at odd hours or during brief windows.

1Ecological Indicators. Active or passive acoustic monitoring? Assessing methods to track anuran communities in tropical savanna wetlands

What the Hardware Looks Like

For much of the twentieth century, acoustic fieldwork meant heavy reel-to-reel tape recorders and a researcher holding a parabolic dish. That changed dramatically with the arrival of small, programmable digital recorders. The most widely adopted is AudioMoth, an open-source device about the size of a credit card with a build cost of roughly $43 per unit. It can sample at rates up to 384 kHz, high enough to capture the ultrasonic calls of bats, and runs on three AA batteries.

2Methods in Ecology and Evolution. AudioMoth: Evaluation of a smart open acoustic device for monitoring biodiversity and the environment

Because AudioMoth is open-source, researchers have adapted it freely. One offshoot, HydroMoth, wraps similar electronics in a waterproof housing for underwater recording at a fraction of the price of traditional hydrophones.

3Remote Sensing in Ecology and Conservation. HydroMoth: Testing a prototype low‐cost acoustic recorder for aquatic environments

Detection algorithms can also be loaded directly onto the device so it records only when it hears something of interest, dramatically cutting power draw and storage needs.

4PubMed Central. Deploying Acoustic Detection Algorithms on Low-Cost, Open-Source Acoustic Sensors for Environmental Monitoring

Tracking Birds, Bats, and Other Terrestrial Wildlife

Terrestrial biodiversity surveys are the most common use of PAM. Arrays of recorders deployed across a landscape can tell you not just which species are present but how active they are and how their behavior changes with habitat quality. One study using automated PAM around hedgerows of different ages in the UK found that bird and bat vocal activity was significantly higher around older, established hedgerows compared to newly planted ones, with young hedgerows (a few years old) showing the strongest gains. That kind of finding helps land managers understand how quickly habitat restoration translates into measurable wildlife recovery.

5Biological Conservation. Using automated passive acoustic monitoring to measure changes in bird and bat vocal activity around hedgerows of different ages

Bat monitoring is a particularly strong use case because bats are nocturnal and echolocate at frequencies humans cannot hear. Autonomous recording units pick up these calls reliably, though automated classification software does occasionally confuse species that sound alike. A study of bat surveys on military lands in Colorado estimated that roughly 1 to 5 percent of detections of certain bat species were misclassified, a useful reminder that the technology works well but is not perfect.

6Methods in Ecology and Evolution. Estimating occupancy from autonomous recording unit data in the presence of misclassifications and detection heterogeneity

Acoustic monitoring even extends to animals whose calls sit below the range of human hearing. African elephants produce low-frequency rumbles, some with fundamental frequencies between 10 and 25 Hz, that travel long distances through both air and ground. Researchers at a zoo near Frankfurt recorded over 1,350 rumbles using co-located seismic and infrasound sensors, demonstrating that you can track elephant vocal behavior non-invasively with ground-mounted equipment.

7PubMed Central. Monitoring Zoo Elephant Rumble Activity Using Combined Seismic and Acoustic Data

Listening to the Ocean

Underwater, sound travels farther and faster than in air, making acoustic monitoring especially powerful for marine ecosystems. Coral reefs are surprisingly noisy places. Healthy reefs hum with the clicks, grunts, and pops of fish and invertebrates, while degraded reefs go comparatively quiet. This difference has turned reef soundscapes into a potential diagnostic tool. A study comparing healthy, restored, and degraded coral reefs found that healthy and restored reefs had a similar diversity of biological sounds, significantly higher than degraded sites.

8Journal of Applied Ecology. The sound of recovery: Coral reef restoration success is detectable in the soundscape

To standardize what “healthy-sounding” means, researchers have developed acoustic indices. The Acoustic Complexity Index (ACI), for instance, measures how variable the sound is across frequencies and time, capturing the richness of a soundscape. On one tropical reef, a low-frequency version of ACI was influenced by fish knocks, herbivorous feeding sounds, and even fishing pressure, explaining about half the variation in the index.

9Ecological Indicators. The utility of different acoustic indicators to describe biological sounds of a coral reef soundscape

In snapping-shrimp-dense waters like those around Singapore, researchers have explored snap rates as an additional metric, since shrimp activity reflects local conditions. Previous work has found positive correlations between sound-pressure levels and live coral cover, and between ACI and fish diversity.

10arXiv. Passive Acoustic–based Composite Indices for Reef Health Monitoring in Noisy Tropical waters

That said, not every acoustic metric tells the same story. The coral restoration study noted that some automated indices replicated the pattern found by manually counting distinct biological sounds, but others did not. This is still a maturing field, and choosing the right metric for the right question matters.

8Journal of Applied Ecology. The sound of recovery: Coral reef restoration success is detectable in the soundscape

Freshwater Environments

Rivers, lakes, and ponds have received far less acoustic attention than forests or oceans, but that is changing. Freshwater environments host a surprising range of sound-producing organisms: fish, frogs, and aquatic insects including beetles, water bugs, dragonfly larvae, and caddisfly larvae.

11Freshwater Biology. Passive acoustic monitoring as a potential tool to survey animal and ecosystem processes in freshwater environments

Physical processes like water flow and sediment transport add their own acoustic signatures, and human activities on or near the water contribute too. PAM in freshwater has already revealed unexpected acoustic diversity, suggesting it could become a practical tool for assessing river and lake health the way it is starting to work for coral reefs.

12PubMed Central. The potential of acoustic monitoring of aquatic insects for freshwater assessment

Soundscape Ecology and the Bigger Picture

Rather than focusing on one species at a time, some researchers treat the entire soundscape as a unit of study. The field of soundscape ecology divides environmental sound into three broad categories: biological sounds (biophony), geophysical sounds like wind and rain (geophony), and human-produced sounds (anthrophony). By analyzing the balance among these three, you can draw inferences about ecosystem health without needing to identify every individual species.

13BioScience. Soundscape Ecology: The Science of Sound in the Landscape

This approach has practical appeal because species-level identification from recordings is labor-intensive and error-prone. A soundscape index can flag that something changed, acting like an early-warning system, even if you do not know exactly which species drove the change. Conservationists can then follow up with targeted surveys in areas where the acoustic signature has shifted.

How AI Is Changing the Analysis

A single AudioMoth recording for a month generates terabytes of audio. No human team can listen to all of it, so machine learning has become essential. Deep-learning models trained on spectrograms (visual representations of sound) can now identify bird species from their songs with accuracy above 96 percent on multiple datasets.

14Ecological Indicators. Deep learning bird song recognition based on MFF-ScSEnet

Noise is the practical bottleneck. Field recordings come packed with wind, rain, traffic rumble, and other interference that can fool classifiers. Dedicated denoising networks can boost the signal-to-noise ratio of bird sounds by more than 10 decibels in most cases, with some scenarios seeing improvements of over 35 decibels.

15Ecological Informatics. Automatic bioacoustics noise reduction method based on a deep feature loss network

That kind of preprocessing makes the downstream species classifiers far more reliable, especially in noisy urban or roadside environments where conservation questions are often most urgent.

Noise Pollution Research

Acoustic monitoring is not just a way to study wildlife; it is also how we study what noise pollution does to wildlife. By recording animals in areas with different levels of human noise and comparing their calls, researchers have found that terrestrial species, especially birds, tend to raise the minimum frequency of their calls in response to noise pollution. They do not generally change how loud or how long they call; instead, they shift the pitch upward, presumably to be heard above the low-frequency rumble of traffic or machinery.

16Journal of Applied Ecology. A meta‐analysis of the influence of anthropogenic noise on terrestrial wildlife communication strategies

The consequences extend beyond communication. A systematic review of noise effects on animal behavior found altered movement patterns, reduced foraging, disrupted reproduction and offspring care, and escape reactions in marine animals. Birds showed changes in foraging and nesting, while laboratory studies on small mammals linked chronic noise exposure to spatial-cognitive impairments and memory loss.

17PubMed Central. Neurobehavioral Alterations from Noise Exposure in Animals: A Systematic Review

Without PAM and related acoustic tools, these patterns would be nearly impossible to quantify at scale.

Industrial Condition Monitoring

Step away from the natural world and acoustic monitoring looks quite different. In industrial settings, the relevant sounds are stress waves: tiny, high-frequency elastic pulses released when materials crack, deform, or rub against each other under load. Sensors that detect these acoustic emissions are attached to machinery, pipelines, or structures to catch problems before they cause failures.

18Expert Systems with Applications. Review of rotating machinery elements condition monitoring using acoustic emission signal

Wind-turbine gearboxes are a good example of why this matters. Bearing failures in wind turbines lead to long downtimes and expensive repairs. Acoustic-emission-based monitoring applied to roller bearings detected surface damage at least 8 hours (about 4 percent of time-to-failure) earlier than traditional vibration sensors. More interesting, the system picked up significant deviations from baseline roughly 130 hours, or about half the remaining life, before test stop, possibly catching sub-surface damage that had not yet reached the bearing surface.

19Wind Energy Science. Condition monitoring of roller bearings using acoustic emission

Structural Health Monitoring

The same principle applies to bridges, buildings, and pressure vessels. Acoustic emission sensors glued to a steel beam can track fatigue cracks in real time. As a crack grows, the acoustic signals it produces change in predictable ways: energy, event count, and signal duration all rise, and they track each other closely enough to serve as reliable indicators of how fast the crack is spreading.

20Measurement. Multi-parameter acoustic emission analysis for fatigue crack evaluation in structural health monitoring

This kind of monitoring can supplement or eventually replace scheduled inspections, catching damage that might not be visible on the surface until it is already serious.

Medical Auscultation Goes Wearable

Listening to the body with a stethoscope is one of the oldest forms of acoustic monitoring. Researchers are now miniaturizing and automating the process. Soft, wearable stethoscope patches can sit on a patient’s chest and continuously stream heart and lung sounds wirelessly to a phone or clinical dashboard. These devices can separate cardiac from respiratory signals even in noisy environments and flag abnormal patterns in real time.

21PubMed Central. Fully portable continuous real-time auscultation with a soft wearable stethoscope designed for automated disease diagnosis

One system, designed as a lung-sound monitoring patch, pairs a wearable sensor with a mobile app that classifies normal versus adventitious (abnormal) breathing sounds by comparing their unique acoustic characteristics. It can distinguish heart sounds from breath sounds and filter out external noise, opening up possibilities for continuous home monitoring of people with chronic respiratory conditions.

22Engineering. A Wearable Stethoscope for Accurate Real-Time Lung Sound Monitoring and Automatic Wheezing Detection Based on an AI Algorithm

Smart Cities and Urban Noise Mapping

Urban planners have adopted acoustic monitoring to map noise pollution across cities. Networks of wireless acoustic sensors can measure sound-pressure levels continuously and feed the data into models that guide traffic management, building codes, and green-space placement. The concept is sometimes framed as part of the Internet of Things infrastructure for smart cities.

23Journal of Sensors. Review of Wireless Acoustic Sensor Networks for Environmental Noise Monitoring in Smart Cities

In Aveiro, Portugal, a project deployed nine noise sensors across the city, each recording continuous sound-pressure levels at five-minute intervals. The resulting data, combined with traffic models, let researchers map noise exposure with far greater spatial and temporal resolution than the occasional spot-check survey that cities traditionally rely on.

24Transportation Research Part D: Transport and Environment. Road traffic noise monitoring in a Smart City: Sensor and Model-Based approach

Volcanoes and Infrasound

Some of the most dramatic acoustic monitoring happens at frequencies well below what any animal can hear. Explosive volcanic eruptions generate infrasound, pressure waves with frequencies below 20 Hz that can travel thousands of kilometers. The International Monitoring System, originally designed to detect nuclear tests, maintains a global network of infrasound arrays that also picks up volcanic activity. Researchers analyzing data from a station in Antarctica, over 1,600 km away, identified infrasound signals from episodes of persistent eruptive activity at Mount Michael in the South Sandwich Islands across a 16-year window. The signal amplitudes were consistent with moderate explosive eruptions, providing evidence of activity at a remote volcano that is difficult to observe any other way.

25Geophysical Research Letters. Long‐Range Multi‐Year Infrasonic Detection of Eruptive Activity at Mount Michael Volcano, South Sandwich Islands

Privacy and the Human Voice Problem

When you leave a microphone in the field to record frogs and birds, it will occasionally pick up people talking. This creates a tension between ecological data collection and individual privacy. Eco-acoustic monitoring is increasingly deployed at huge scales, yet relatively little thought has historically been given to the privacy implications of inadvertently recorded human speech.

26bioRxiv. Automated speech detection in eco-acoustic data enables privacy protection and human disturbance quantification

Automated speech-detection algorithms can now flag and strip segments of audio that contain human voices before the data are shared or archived. The same detection capability has a bonus scientific use: quantifying the level of human disturbance at a site by measuring how often speech is present. Still, as monitoring networks grow denser and recorders become cheaper, the ethics of ambient listening in public and semi-public spaces will need ongoing attention from both researchers and regulators. Some jurisdictions already treat ambient audio recording differently from, say, camera-trap photography, and the legal landscape is patchy enough that project-by-project review remains the safest approach.